Author

Jen-Wen Lin

Date of Award

2006

Degree Type

Thesis

Degree Name

Doctor of Philosophy

Program

Statistics and Actuarial Sciences

Supervisor

Dr. A. Ian MacLeod

Second Advisor

Dr. Guangyong Zou

Abstract

Portmanteau tests are standard diagnostic tools under the Box-Jenkins framework and are widely used in practice. This thesis developed new portmanteau tests for diagnostic checking several time series models, such as multivariate time series with autoregressive conditional heterscedastic errors and autoregressive models with stable Paretian errors. Using a proposed Monte-Carlo test procedure, I found that the size distortion of the proposed tests is negligible and that the proposed tests are usually more powerful than usual portmanteau tests. A nonparametric bootstrap test procedure is also used to check randomness of a time series. It is found that the size distortion of portmanteau tests with the boot strap test procedure is negligible except for highly skewed distributions and extremely heavy tailed distributions. In such asymmetric distributions, the use of a Box-Cox transformation on the data prior to bootstrapping eliminates the size distortion prob lem. In all, in this thesis, new portmanteau tests based on the generalized variance of standardized residuals were developed and found to be more powerful empirically than usual portmanteau tests. It is also found that the use of the Monte-Carlo test or bootstrap test procedure eliminates the size distortion problem which is the common criticism on the use of portmanteau tests.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.